Case Study: How Bolt cut mean time to detect by 99% with Sentry
Key results
The challenge
Bolt, an AI-native application builder launched by StackBlitz in 2024, hit observability gaps after rapid growth to 250,000 weekly active users within two months. Its legacy monitoring stack had been built for a Rails backend, but Bolt ran on Remix and Cloudflare Workers, leaving bug triage overwhelming.
The solution
Bolt adopted Sentry for first-class support of its modern JavaScript stack, gaining visibility into AI model workloads and catching token-consumption anomalies that affected unit economics. The team also used Seer, Sentry's AI debugger, to automate initial error triage.
The results, in context
Bolt reported a 99% reduction in mean time to detect (from days to hours) and 18% fewer errors in a key workflow, after scaling to 250,000 weekly active users in its first two months. The figures are quoted from Sentry's published customer story and dated to when sourced, July 24, 2026.